- Home
- Skills
- DevOps & Cloud
- Log Analyzer: Gigabytes of Logs to the Real Error
More screenshots
Works with the AI Agents you already use
Log Analyzer: Gigabytes of Logs to the Real Error
Turn a huge log file into a 2 KB summary with the real error on top: templates, counts, folded stack traces and what is new since the last good run.
$12
About this skill
Your log is 400 MB. The agent reads the first 200 lines, greps for "error", gets thousands of hits of the usual noise, and guesses.
On a public 743 MB, 4.7-million-line supercomputer log, Log Distiller's summary is 2,033 bytes, and 100% of the 348,460 labelled alert lines land in the templates it flags as problems.
What the tool does (one Python script, standard library only, runs in seconds to a minute):
- Reads everything, in one streaming pass. Files, folders, .gz, stdin; UTF-16 PowerShell output and CRLF too. Memory stays bounded; big files run on all cores.
- Folds stack traces into the log line that caused them and names each by root cause and your own frame:
requests.exceptions.ReadTimeout @ charge.py:88 in charge_card(JavaCaused by, Python, Node, .NET, Go panics). - Groups lines into templates (ids, IPs, numbers, paths masked) with count, first/last time and first/last line number, so the agent opens exactly the right lines.
- Ranks rare, severe, recent problems first, lists the loudest error floods separately, and keeps the output to about 2 KB.
- Diffs against a good run with
--baseline: what is NEW, what SPIKED, what is GONE. - Drills down with
--show T12: the real records behind one template.
Benchmark on public loghub datasets (same error injected 3 times, then found among real noise):
| Log | Size | Summary | Rank of the injected error | Rank among NEW vs a baseline |
|---|---|---|---|---|
| Zookeeper | 10.4 MB | 1,714 B | 1 of 21 problems | 1 of 3 |
| OpenSSH | 73.4 MB | 2,104 B | 3 of 29 problems | 2 of 5 |
| Apache | 5.1 MB | 2,161 B | 1 of 19 problems | 1 of 2 |
| BGL (labelled) | 743 MB | 2,033 B | 6 of the top 8 problems are real alert types | 45.6 s on 8 cores |
Exact commands, dataset checksums and the misses are in the skill's references/how-it-works.md.
What it does not do: it finds what was logged. A step that only got slower, with no new message, is invisible to it (on the OpenStack dataset, whose anomalies are latency, it found nothing new). Templates are heuristic, so read the three sections before concluding. It does not fix code or watch logs live.
Works in Claude Code, Cursor, Codex CLI, Copilot and any SKILL.md agent.
Privacy: runs locally on the files you name, no network, no telemetry. Summaries quote real log lines, so review them before sharing.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 27 days ago